Does quantum mechanics violate Bell's inequality?

The EPR paradox is known as an interpretive problem, as well as a technical discovery in quantum mechanics. It defined the basic features of two-quantum entanglement, as needed to study the relationships between two non-commuting variables. In contrast, four variables are observed in a typical Bell experiment. This is no longer the same problem. The full complexity of this process can only be captured by the analysis of four-quantum entanglement. Indeed, a new paradox emerges in this context, with straightforward consequences. Quantum mechanics cannot violate CHSH-type inequalities, in the same sense in which local realism cannot do it. The solution is to assume that quantum correlations do not work at the level of full populations, when violations occur, but only apply to incompatible slices of input beams.